charles%hplcyy@HPLABS.HP.COM (Charles Young) (10/19/88)
I have a simple request to make about changing the screen saver behaviour. Instead of the screen going blank when there are no input for a while, why not change it so that it only goes blank when there are no "drawing" on screen. With this small change, I can watch the machine slave over my program generating output without me slaving over the mouse every couple of minutes! Thanks
charles%hplcyy@HPLABS.HP.COM (10/20/88)
The objection that xclock and xload will defeat the screen saver does not really make sense because most people has a "xlock" program that serves as a screen saver and security when they are really away from the display. It is quite annoying when my client program is happily updating the display and all of a sudden, the display goes blank and you have to move the mouse or something to redraw the screen. Particularly so when certain information may be lost duing the blank out time. Charles (I don't want to hack my server) Young
kek@DINORAH.WUSTL.EDU (Ken Krippner) (10/20/88)
another vote for a screen saver that awakens for screen writing as well as mouse or key activity.
swick@ATHENA.MIT.EDU (Ralph R. Swick) (10/20/88)
> Charles (I don't want to hack my server) Young
So don't hack your server; just do 'xset s 3600'. (If you can keep
your fingers off the keyboard/pointer for an hour and your program
is still drawing they maybe you need to modify something else :-)naughton%wind@Sun.COM (Patrick Naughton) (10/20/88)
It sounds like the people with xclock and other continuous output X
clients who want the screen saver to reset on screen output simply want
to disable the screen saver entirely. If they use xlock(1) as their
screen saver when they really want one, they should just use xset(1) to
disable the screen saver:
% xset -s off
or a simple program that does this:
. . .
XSetScreenSaver(dpy, 0, 0, 0, 0); /* disable screen saver */
. . .
-Patrick
p.s. The new xlock has a -saver option to just put up the graphical display
and not lock the display.
______________________________________________________________________
Patrick J. Naughton ARPA: naughton@Sun.COM
Window Systems Group UUCP: ...!sun!naughton
Sun Microsystems, Inc. AT&T: (415) 336 - 1080eirik@tekcrl.TEK.COM (Eirik Fuller) (10/21/88)
In article <8810191538.AA12619@hplcyy.HPL.HP.COM> charles%hplcyy@HPLABS.HP.COM (Charles Young) writes:
) ...
) Instead of the screen going blank when there are no input for a while, why
) not change it so that it only goes blank when there are no "drawing" ...
Please no; they made this mistake in X10. You evidently don't use a clock.
I like the approach some mac screen savers take, in which there are
two special rectangles, one for blank right away, another for blank
never. For long output only programs, you'd just park the cursor in
the magic spot that disables screen blanking. Other interfaces to
this would be acceptable too, like menus perhaps.eichin@ATHENA.MIT.EDU (Mark W. Eichin) (10/21/88)
>>Date: 21 Oct 88 07:21:36 GMT >>From: tektronix!tekcrl!eirik@bloom-beacon.mit.edu (Eirik Fuller) >>In article <8810191538.AA12619@hplcyy.HPL.HP.COM> charles%hplcyy@HPLABS.HP.COM (Charles Young) writes: >>) Instead of the screen going blank when there are no input for a while, why >>) not change it so that it only goes blank when there are no "drawing" ... >>Please no; they made this mistake in X10. You evidently don't use a clock. The way I heard the story was that the Athena needed something to get rid of the screensaver, when we had VS100's running X10 that were used by students registering for accounts -- otherwise non-technical students would assume they were broken. So, Tony DellaFera (DEC employee @Athena) wrote an early version of xclock. >>I like the approach some mac screen savers take, in which there are >>two special rectangles, one for blank right away, another for blank >>never. For long output only programs, you'd just park the cursor in >>the magic spot that disables screen blanking. Other interfaces to >>this would be acceptable too, like menus perhaps. `xmoire' (posted to comp.sources.x by me) is such a MAC screen saver; it was a look-and-feel reconstruction, based on staring at the `moire' screen saver on our Mac 2 for far too long. It was a quick hack; as it turns out, I never use it. It has the upper-left corner `activate' square. A more popular alternative seems to be the screen-lock program we use here, which was posted a long time ago (it had been ported from X10) -- it has a little square which you click in to pop a `save/lock/exit' menu from. It then copies the background over the screen, and bounces the little square (with a timeclock and bitmap in it) around the screen. Any button- or key- press puts up a box with a password prompt. ------------ Keep in mind my proposal from a few months ago --- someone should write a `screensaver redirect/notify' extension, such that there would exist a portable way for a random client to know that the X server thought the user was idle. As this discussion aptly proves, if you don't let the customer choose the behavior, the customer will flame. Mark Eichin <eichin@athena.mit.edu> SIPB Member & Project Athena ``Watchmaker''
august@ole.UUCP (Augustine Chan) (10/25/88)
In article <8810201154.AA15644@LYRE.MIT.EDU> swick@ATHENA.MIT.EDU (Ralph R. Swick) writes: >> Charles (I don't want to hack my server) Young > >So don't hack your server; just do 'xset s 3600'. (If you can keep >your fingers off the keyboard/pointer for an hour and your program >is still drawing they maybe you need to modify something else :-) When your client program is graphics intensive, e.g., a VLSI layout editor displaying a microprocessor, the displaying could take more than an hour. Part of the reason it can take this long is the unimpressive graphics speed of X11R2 on popular platforms such as the Sun and the Apollo. The graphics speed of our applications is about 3 to 4 times slower on Apollo(DN3000)-X11 than on Apollo's GPR graphics interface or on Sun3-X10. The speed on Sun3-X11 is another 3 to 4 times slower than on Apollo-X11. Something that takes 5 minutes to display on Apollo-GPR or Sun3-X10 can take up to an hour on the Sun3-X11 server. Until there is satisfactory graphics speed on X11, I'll always use 'xset s off'. Augustine Chan
raj@PARIS.ICS.UCI.EDU (Richard Johnson) (10/29/88)
It seemed fairly obvious to me back under X10, and it still seems obvious, that every client should be able to specify to the server whether or not it's output is to be considered "important". Clients' outputs would all be considered "important" unless they specifically told the server it wasn't. This way as long as the server is getting output from clients who have told it their output is "unimportant", and hasn't gotten any output from "important" clients, it could go ahead and enter screensaver. Then we just need versions of xclock, xload, etc. which specify their output as being unimportant.